Moist dynamics of tropical convection zones in monsoons...

Post on 27-Jun-2020

1 views 0 download

Transcript of Moist dynamics of tropical convection zones in monsoons...

Moist dynamics of tropical convection zones inMoist dynamics of tropical convection zones in monsoons, monsoons, teleconnectionsteleconnections and global warming and global warming

• Leveraging a theoretical framework from convective quasi-equilibrium, moist static energy budget, + ….

• Monsoon aspect: what limits seasonal movement of deepconvection zones over continents?

• Tropical regional precipitation anomalies associated withchanges in deep convection zones: including droughtregions in both El Niño & global warming cases.Mechanisms? Similarities?

J. David Neelin*J. David Neelin*

*Dept. of Atmospheric Sciences & Inst. of Geophysics and Planetary Physics, U.C.L.A.

Collaborators: Collaborators: ChiaChia Chou**, Chou**, Hui Hui Su*, Katrina Hales*, Chris Holloway*Su*, Katrina Hales*, Chris Holloway*

**Inst. of Earth Sciences, Academia Sinica, Taiwan

Temperature Temperature TT and Moisture and Moisture qq equations equations

Quasi-equilibrium schemesQuasi-equilibrium schemes

Posit that bulk effects of convection tend to establish statisticalequilibrium among buoyancy-related fieldsApproach here depends on convection tending to constrainvertical structure of temperature field.For now: Smoothly posed convective adjustment

Convective heating: (Betts 1986; Betts & Miller 1986)Qc = (Tc - T)/τc

τc time scale of convective adjustmentTc convective profile; may interact with hbhb atm boundary layer (ABL) moist static energyTc typically moist adiabat or closely relatedCan be expanded about a reference state, Tr

c

Tc = Trc (p) + A(p) T1

c + higher order

A(p) vertical dependence of the moist adiabat perturbation T1

c incl. ABL adjustment by downdrafts to satisfy energy constraint Tends to reduce CAPE (convective available potential energy)

Manabe et al 1965; Arakawa & Schubert 1974; Moorthi & Suarez 1992; Randall & Pan 1993; …

Vertical structure of temperature variationsVertical structure of temperature variations

• Regressions of Temp. on 850-200 hPa Ave. Temp. (squares)• Same for reversible moist adiabats over similar range (grey)

Courtesy, C. Holloway

Daily, COARE Domain Monthly, tropical avg.

Rawinsonde (4 mon) NCEP (2 yr) NCEP (24 yr)

See also, e.g., Xu and Emanuel 1989; Fu et al 1994; Brown and Bretherton 1997; Sobel et al 2004

[Ciesielski et al (2003) data]

QE analyticQE analyticAnalytical solution under quasi-equilibriumAnalytical solution under quasi-equilibriumconvective constraintsconvective constraints

If T constrained to be close to QE temp Tc

baroclinic pressure gradients have strongly constrainedvertical structure

Primitive equations, momentum + hydrostatic:

( ) cc

r

cTpATTT1

+!!

( ) ! "+"#=$++%op

p

omt pTdvfvD &' lnk

( ) o

c

p

p

TpdpAo

!" #+#$% & 1ln

Analytic solution in deep convective regions (cont.)Analytic solution in deep convective regions (cont.)

Vertical structure of baroclinic pressure gradients⇒ structure of baroclinic wind V1. With barotropiccomponent ⇒ v = vo(x,y,p,t) + V1(p)v1(x,y,t)

Continuity eqn. ⇒ ω = Ω1(p) ∇·v1

The moist static energy eqn. becomes

(∂t + D)(T + q) + M∇·v1 = Fnet

where M is the gross moist stability M = 〈Ω∂ph〉

NB: Have not yet used convective closure on moisture.

^ ^

Neelin 1997; Yu et al 1998; Neelin and Zeng 2000

Moist convection interacting with large-scale dynamicsMoist convection interacting with large-scale dynamics

•• Convective Quasi- Convective Quasi- Equilibrium: Equilibrium: Fast convective motions reduce Convective Available Potential Energy (CAPE) Constrains temperature through deep column Baroclinic pressure gradients

•• Gross moist stability Gross moist stability at large scales at large scales

Refs: Arakawa & Schubert 1974;Emanuel et al 1994; Neelin & Yu 1994;Brown & Bretherton 1997; Neelin & Zeng2000

GalerkinGalerkin expansion expansion

Have approx. analytic solution for convective regions:Extend to full nonlinearity, non-convective regions,…Use analytic solutions for leading basis functions inGalerkin expansion in vertical

•Horizontal gradients of T matter; specify reference stateTr(p) to improve accuracy.•Simplest case: 1 basis function in T, q

Extra basis function for external mode ⇒ 2 in v

,),,()()(1

!=

++=K

k

Rkkr TtyxTpapTT

Neelin and Zeng (2000)

Quasi-equilibrium Tropical circulation model:Quasi-equilibrium Tropical circulation model:

• Primitive equations projected onto vertical basis functions fromconvective quasi-equilibrium analytical solutions

• for Betts-Miller (1986) convective scheme, accurate verticalstructure in deep convective regions for low vertical resolution

• GCM-like parameters but easier to analyzeRadiation/cloud parameterization:Radiation/cloud parameterization:• Longwave and shortwave schemes simplified from GCM

schemes (Harshvardhan et al. 1987, Fu and Liou 1993)• deep convective cloud, CsCc fraction param. on precipSimple land model:Simple land model:• 1 soil moisture layer; evapotranspiration with stomatal/root

resistance dep. on surface type (e.g., forest, desert, grassland)• low heat capacity; Darnell et al 1992 albedo

• http://www.atmos.ucla.edu/~csi/QTCM

• Seasonal movement of deep convection zones over continentsdeep convection zones over continents• Dynamical mechanisms mediating land-ocean contrast?• Given the large insolation extending poleward over continents, why

do deep convection zones not extend farther poleward?• Do mechanisms affecting convection zones differ from continent to

continent?• QTCM coupled to a mixed-layer ocean and simple land model• Focus on dynamical aspects, less on surface type• No-topography case emphasizes ocean-land contrast

Dynamics of summer monsoon convective zonesDynamics of summer monsoon convective zones

Wind-based vs. Wind-based vs. precipprecip. regime-based monsoon definitions. regime-based monsoon definitions

Freq. of surface wind direction; Khromov (1957); from Ramage (1971)

Pronounced summer precip maximum; sfc low; upper level high andoutflow, … (e.g., Tang and Reiter 1984; Barlow et al 1998; Higgins et al 1998;Zhou and Lau 1998; Yu and Wallace 2000; …)

i.e., characteristics of seasonal arrival of deep convection zone

Seasonal precipitation minus Annual AverageSeasonal precipitation minus Annual Average

DJF ave. – Annual ave.

JJA ave. – Annual ave.

mm/day

Latitude-height cross section at 90E from Bay of BengalLatitude-height cross section at 90E from Bay of Bengalacross Tibetan Plateau across Tibetan Plateau (shaded regions are rising motion)(shaded regions are rising motion)

From Yanai et al. (1992)

precFnet Observed climatology JanuaryObserved climatology January

Net Flux into theatmosphere

Precipitation

precFnet Observed climatology JulyObserved climatology July

Net Flux into theatmosphere

Precipitation

Precipitation Net flux into atmosphere

Low-level wind Upper-level wind

QTCM climatology JulyQTCM climatology July(coupled to a mixed-layer ocean)(coupled to a mixed-layer ocean)

QTCM1V2.2

Precipitation Net flux into atmosphere

Low-level wind Upper-level wind

4panel Observed climatology JulyObserved climatology July

Ventilation and the Ventilation and the ““interactiveinteractiveRodwell-HoskinsRodwell-Hoskins”” mechanism mechanism

Chou, Neelin and Su 2001

The The ““ventilation mechanismventilation mechanism””

• import of low moist static energy air from oceanwhere heat storage opposes summer warming

• oceanic air: cooler and moisture is lower thanconvection threshold over warm continent

• import to continents by wind (including upperlevel jets) via advection terms in temperature andmoisture equations

Chou, Neelin and Su 2001

The The ““interactiveinteractive Rodwell Rodwell-Hoskins mechanism-Hoskins mechanism””

• Rodwell and Hoskins (1996): imposed convectiveheating in Asia gives Rossby wave descent pattern towest, enhancing deserts.

• when convection is interactive: associated flow feedsback on heating, creating characteristicconvection/dry region pattern» we emphasize feedback

(convection ⇔ baroclinic Rossby wave dynamics),hence:

» “interactive Rodwell-Hoskins” (IRH) mechanism

Chou, Neelin and Su 2001

ControlSaturated soil moisture over

South American region

No ventilation: v•∇q, v•∇T set tozero over South American region

No ventilation and no β-effect:f = constant in South American

region (9S-56S - 70W-20W)

South American region case (observed albedo) JanSouth American region case (observed albedo) JanPrecipitationPrecipitation

Chou and Neelin 2001

Refinement of experimental designRefinement of experimental design

• Irrotational (purely divergent) wind component vχ• Non-divergent wind component vψ “No ventilation” = suppress vψ •∇T, vψ •∇q Retains conservation property:

Domain

(vχ •∇q + q ∇ • v)dA = 0since ∇ • vψ= 0

1. Consistent treatment of vχ :

Asian region case Asian region case –– July July (1) V2.3b

Control Saturated soil moisture

No ventilation: v•∇q, v•∇T set to zero No β-effect: f = constant

PrecipitationPrecipitationAsian region case Asian region case –– July July (1) V2.3b

Chou and Neelin 2003QTCM v2.3

Saturated soil moisture

North American region case North American region case 1 V2.3b

Control

No ventilation: v•∇q, v•∇T set to zero No β-effect: f = constant in region

July PrecipitationJuly Precipitation

Chou and Neelin 2003

North American region case North American region case 3 V2.3b

No T ventilation No q ventilation

July PrecipitationJuly Precipitation

Chou and Neelin 2003

Control No ventilation: v•∇q, v•∇T set to zero

Saturated soil moisture

No ventilation: v•∇q, v•∇T set to zero No ventilation and no β-effect:

African region case (observed African region case (observed albedoalbedo) July ) July V2.3b

ControlPrecipitationPrecipitation

Chou and Neelin 2003(red contour: albedo = 0.3)

Saturated soil moisture

No ventilation: v•∇q, v•∇T set to zero No ventilation and no β-effect:

African region constant African region constant albedoalbedo case ( case (0.26 over Africa0.26 over Africa) July ) July V2.3b

ControlPrecipitationPrecipitation

Chou and Neelin 2003

Mechanisms affecting convective zones (S. American case)Mechanisms affecting convective zones (S. American case)

Ocean heat transportout of the tropics

Ventilation and theinteractive

Rodwell-Hoskins mechanism

African northern summer monsoon climatologyAfrican northern summer monsoon climatology

• Albedo is leadingeffect on polewardextent of convection

• Dynamicalmechanisms: affect margin ofconvective zone take over if albedogradient is flattened

African monsoon & mid-Holocene (6kaBP) green SaharaAfrican monsoon & mid-Holocene (6kaBP) green Sahara

• Paleo Model Intercomp Proj (PMIP): orbital parameterchange => small monsoon boundary shift (Joussaume et al1999) rel to pollen data (Jolly et al 1998; Prentice & Webb 1998)

• LMD and ECHAM GCMs with same interactive veg modelyield different results (DeNoblet et al 2000)

• QTCM expts with grass-like albedo over Sahara (Su et al, inprep) & interactive vegetation (Hales et al, in prep)

• Ventilation opposes rise of moisture; may not increaseenough to reach increase in trop temp.

• Reduced ventilation expts: e.g. reduce advection affecting qanomalies rel. to control; or reduce ventilation in control and6kaBP cases

African mid-Holocene 6kabp monsoonAfrican mid-Holocene 6kabp monsoonSchematic of QTCM - Schematic of QTCM - SVeg exptsSVeg expts..

• Tropical regional precipitation anomaliesassociated with changes in deep convection zones:including drought regions in both El Niño &global warming cases. Mechanisms? Similarities?

Mechanisms for regional precipitation anomalies inMechanisms for regional precipitation anomalies inteleconnectionsteleconnections and and

global warmingglobal warming

www.atmos.ucla.edu/~csi

Observed anomalies during July-Nov 1997Observed anomalies during July-Nov 1997

Precipitation(mm/day)

TroposphericTemperature

Neelin et al 2003; Neelin and Su 2004 subm

QTCM anomalies forced by Pacific positive SSTQTCM anomalies forced by Pacific positive SSTanomalies July-Nov 1997anomalies July-Nov 1997

TroposphericTemperature

Precipitation(mm/day)

QTCM POSPAC-FluxesQTCM POSPAC-Fluxes

Surface temperature

Net surface flux

Net flux intoatmospheric column

QTCM July-Nov 1997: Anomaly budget contributionsQTCM July-Nov 1997: Anomaly budget contributions11

Moistureconvergence

Mq∇·v(by divergent flow)

Moist staticenergy

divergence*

M ∇·v

* Gross moist stability M=Ms–Mq is an effective stability that includes partial cancellation of adiabatic cooling by diabatic heating

QTCM July-Nov 1997: Anomaly budget contributionsQTCM July-Nov 1997: Anomaly budget contributions22

Radiative coolinganom. (top of

atmosphere) dueto temp. anom.

Mean windadvection of

moisture anomalyv·∇q'

ENSO ENSO teleconnectionsteleconnections to regional to regional precipprecip. anomalies. anomalies

Su & Neelin, 2002

ENSO ENSO teleconnectionsteleconnections to regional to regional precipprecip. anomalies. anomalies

• a small zoo of mechanisms with moist convective and cloudradiative feedbacks

Zeng & Neelin 1999; Giannini et al 2001; Su et al 2001; Bretherton & Sobel 2002;Chiang and Sobel 2002; Chiang et al 2002; Su and Neelin 2002; Neelin et al 2003;Neelin and Su 2004 subm…

The The ““upped-anteupped-ante”” mechanism mechanism22

Margin ofconvective zonewith v inward

from dry region

Neelin, Chou & Su, 2003

QTCM experiments suppressing upped-ante mechanismQTCM experiments suppressing upped-ante mechanism

Control

(v · ∇q)'

T' contribution toCAPE

Anomaly ( )' termsuppressed in region:

Precipitation Anomalies

Other mechanismsOther mechanisms

• MMoist SStatic EEnergy transport by divergent flow ≈ M ∇·v• GGross MMoist SStability MM = MMss- MMqq, (Mq inc. with moisture)

Perturbation MSEMSE budget + ocean mixed layer / landocean mixed layer / landM ∇·v' = –M' ∇·v – (v·∇q)' – cc∂∂ttTT '' + F net' + (v·∇T)' …

Yields precip anoms as T' ⇒ q' ⇒ ∇q' , M' ; v' , q' ⇒ E' etc.

P' ≈ –(v·∇q)' + ∇·v(–M' ) – cc∂∂ttTT '' + …

tops

GMSGMS multiplier effect

MqM [ ]

Upped-ante

SST disequilibrium

Anomalous GMS Rad cooling, (v·∇T)' ocean transp, …

ss

Mixed layer ocean in Atlantic: 1Mixed layer ocean in Atlantic: 1stst year year vsvs equilibrium equilibrium

•Long termequilibrium withEl Nino SST anomartificiallysustained

•Positive 1997 ElNino Jul.-Nov.SST anom inPacific: tropicalAtlantic 50m MLSST is adjusting

Precipitationanomalies

(anomalies relative to climatology of ML Atlantic, clim. SST elsewhere)

Global Warming case: GCM Global Warming case: GCM PrecipPrecip. . AnomAnom..

DJF precip. anom.Three GCM

Greenhouse gasscenarios for

2070-2090rel. to 1961-1990 clim

QTCM doubled COQTCM doubled CO22 experiments experimentsQfluxQflux mixed-layer ocean mixed-layer ocean

Dec - FebPrecip change

Dec - FebQTCM Precip

climatology

Neelin et al 2003; Chou & Neelin 2004

Mq' ∇·v

Anomalous moistureconvergence due to

moisture anom. q'

QTCM doubled COQTCM doubled CO22 experiments experimentsMoisture budget contributionsMoisture budget contributions 11

(v·∇q)'

Anomalous moistureadvection

QTCM doubled COQTCM doubled CO22 experiments experimentsMoisture budget contributionsMoisture budget contributions 22

Mq ∇·v'

Anomalous moistureconvergence due to

anomalous divergence(GMS multiplier effect

feedback )

The The ““upped-anteupped-ante”” mechanism mechanism11

Margin ofconvective zonewith v inward

from dry region

Neelin, Chou & Su, 2003

QTCM 2xCOQTCM 2xCO22 ExptExpt. suppressing change in moisture. suppressing change in moistureadvectionadvection

Experiment2xCO2 Precip. change

(mm/day)

Control2xCO2 Precip. change

(testing the upped-ante mechanism)

Response to imposed T change in CAPEResponse to imposed T change in CAPE

• T' =1.5 C added to temperature only inside convection scheme• Mimicks 2xCO2 moisture and regional precip response• DJF Precip (W/m2), surface temp, moisture (K), tropospheric mean temp

Anomalous Gross Moist Stability (MAnomalous Gross Moist Stability (M'') mechanism) mechanism

•• MMoist SStatic EEnergy transport by divergent flow ≈ M ∇·v• M = Ms- Mq

• M ∇·v' + M' ∇·v = F'net - (v·∇q)' + …

• P' ≈ ∇·v(-M')

• Mechanism increases convergence & precip. in strongconvergence zones: “rich-get-richer”

may partially compensate if cloud top risesincreases with increasing moisture, tends to reduce M

reducedincreases to compensate

MMq

QTCM 2xCOQTCM 2xCO22 ExptExpt. suppressing change in gross moist. suppressing change in gross moiststability, Mstability, M

Experiment2xCO2 Precip. change

(mm/day)

Control2xCO2 Precip. change

(testing the M' mechanism)

Summary: Regional Summary: Regional precipprecip. . TeleconnTeleconn./global warming./global warming• tropical regional precipitation anomalies in ENSO teleconnection

case ⇒ a handful of contributing mechanisms• 2XCO2 case, mixed-layer ocean case ⇒ set of mechanisms with

some cross-over to ENSO case• the "upped-ante mechanism": substantial negative precip. anomaly regions in both cases drought occurs at margins of convection zones with climatologicalwind inflow from dry zone into convection zone

• the "anomalous gross moist stability (M') mechanism": contributes positive precipitation changes in strong precipitationregions in global warming case (but theory for M' very poor)

• Surface heat fluxes & SST′ important only when SST is indisequilibrium (∂t SST or ocean transport anom.)

Regional Regional precipprecip. . anomanom. relation to monsoon case. relation to monsoon case

• role of QE mediation: moisture rel to T of free troposphere vsimported q

• Fnet: favorable but not sufficient in monsoon; small in land anoms ocean heat storage & transport ⇒ ocean land contrast in monsoon,contrib. some Atl. negative precip. anoms. in ENSO case

• vψ •∇(q, T) terms: strong impact on precip/precip anomalies dueto role in moist static energy budget partial cancellation of adiab. cooling & diab. heating ⇒ small MMq/M large ⇒ GMS multiplier effect oppose precip in parts of trop convg zones & limit poleward extent role in upped-ante mechanism; involves QE to not QE transition

Last SlideLast Slide

• Title page

• ENSO teleconnections section

• Global warming case

• End show

Neelin 1997

Model Summary Model Summary –– QTCM equations QTCM equations

∂tv1 + DV1(v0,v1) + fk x v1 = -κ∇T1 - stress

∂tv0= … (barotropic component)

a1(∂t + DT1)T1 + MS1∇·v1 = 〈Qc〉 + Rad + H

b1(∂t + Dq1)q1 + Mq1∇·v1 = 〈Qq〉 + E

Moisture sink and convective heating

−〈Qq〉 = 〈Qc〉 = εc(q1 – T1)

^

^

Zeng, Neelin and Chou 2000 QTCM1 V2.0 clrad1

ENSO Composite (DJF)ENSO Composite (DJF)

QTCM climatology JanuaryQTCM climatology January(coupled to a mixed-layer ocean)(coupled to a mixed-layer ocean)

Precipitation Net flux into atmosphere

Low-level wind Upper-level wind

QTCM1V2.2

4panel Observed climatology JanuaryObserved climatology January

Precipitation Net flux into atmosphere

Low-level wind Upper-level wind

QTCM experiments suppressing various mechanismsQTCM experiments suppressing various mechanisms

T' radiative effects

(v · ∇T)'

(surface stress)'

Precipitation AnomaliesAnomaly ( )' term

suppressed in region: